Crazyflie Autonomous Navigation

This Python script controls a Crazyflie quadrotor in Webots simulator, generating a trajectory using the A* path planning algorithm and following it using PID control. Additionally, the script utilizes YOLO for real-time person recognition, causing the quadrotor to stop upon detection.

Authors

• @josemaese
• @santiagomoreno0123
• Miguel Ballesteros Delgado
• @javgilavi

Dependencies

• controller, Keyboard, and Supervisor modules from Webots.
• PID controllers (pid_velocity_fixed_height_controller and QuadrotorController) for stability and navigation.
• Computer Vision functionalities with OpenCV (cv2) for person recognition using YOLO.
• Other libraries include numpy, math, threading, csv, PIL, matplotlib, and custom modules like ruta_predefinida, evita_obstaculos, astar3D, etc.

Usage

1. Install Webots.
2. Place the script in the Webots project directory.
3. Run the Webots simulator with the specified Crazyflie model.

Controls

• UP, DOWN, RIGHT, LEFT: Move in the horizontal plane.
• Q, E: Rotate around yaw.
• W, S: Go up and down.
• A: Enable autonomous mode.
• D: Disable autonomous mode.
• 1, 2, 3: Set predefined targets.
• I: Export data to CSV and generate graphs.

Autonomous Mode

• The script generates a trajectory using the A* path planning algorithm and follows it autonomously.
• YOLO-based person recognition triggers a stop action.

Vision System

• YOLO (You Only Look Once) is used for real-time person detection.
• Camera feed is processed to recognize and react to the presence of people.

Additional Features

• Data logging: The script logs positional data and target information into CSV files for further analysis.